TF 2.0 is out! Get hands-on practice at TF World, Oct 28-31. Use code TF20 for 20% off select passes. Register now

tf.contrib.gan.get_sequential_train_steps

View source on GitHub

Returns a thin wrapper around slim.learning.train_step, for GANs.

tf.contrib.gan.get_sequential_train_steps(train_steps=namedtuples.GANTrainSteps(1, 1))

This function is to provide support for the Supervisor. For new code, please use MonitoredSession and get_sequential_train_hooks.

Args:

  • train_steps: A GANTrainSteps tuple that determines how many generator and discriminator training steps to take.

Returns:

A function that can be used for train_step_fn for GANs.